My review of OpenAI's DALL-E 2 images generator (as a clothes and NFT creator)
OpenAI created a strong buzz around Dall-E 2 revealing impressive pictures, making people wait for access, and feeding debates about what people could do wrong with that kind of powerful tools.
As an artist playing with artificial intelligence since late 2021, I was both excited and frightful about the release:
Will AI digital art generation become mainstream?
How will I preserve a strong artist identity if anyone could generate astronaut dogs?
Will it erase all my previous efforts to understand how to be understand by previous models?
Here's some responses:
1- DALL-E 2's results are not so impressive
fails et trucs moches ; résultats ennuyants car trop réussi
prompt='a dragon with butterfly wings is on a lotus flower on a sea, digital art 4K trending on artstation'
2- DALL-E 2 is really great at understanding my composition desire
prompt='rainforest jungle scene with a parrot in the sky, creepers and snakes on trees, a haystack, and a river downside with eels and lampreys in it, 4K trending on artstation'
3- New AI models ain't threats to old ones : they are complementary
I can compose the scene with DALL-E 2 then draw on it with other AI
I can compose by hand then ask DALL-E 2 to draw on it
TIPS more please!
I live on the road (by bike) with almost nothing!..